Two of the most fundamental mechanisms in biological vision are attention and plasticity. Attention actively selects and enhances visual information most relevant to behavior. Neural plasticity enables the visual system to benefit from perceptual experience. The amount of visual information to learn is infinite, however, so top-down control mechanisms must regulate learning to maintain a balance of plasticity and stability in neural circuitry. This proposal explores how attention modulates perceptual learning according to behavioral goals, such that only attended information will induce durable changes in neural responses throughout multiple stages along the visual pathway. Surprisingly very little neuroimaging work has been devoted to this basic hypothesis. The proposed studies use functional magnetic resonance imaging (fMRI) to investigate how attention modulates neural plasticity, as revealed by changes in the blood oxygen level dependent (BOLD) signal, an indirect measure of neural activity that can be obtained non-invasively from the human brain. Study 1 will measure plasticity along the visual processing stream to reveal the stage at which attention begins to suppress short-term and long-term cortical plasticity to unattended images. Study 2 will test the hypothesis that cortical plasticity in visual areas is guided by working memory representations of target visual events. Study 3 will examine the relationship between perceptual learning and the fMRI BOLD signal. Altogether, the findings should 1) reveal where attention begins to modulate cortical plasticity to recurrent visual features and images, 2) clarify the function of top-down attentional mechanisms in regulating plasticity, and 3) tighten the complex relationship between learning-related changes in fMRI BOLD signal and enhanced perceptual performance across a wide range of visual tasks and cortical mechanisms. Novel data from whole-brain imaging may guide future electrophysiological or neuropsychological research on top-down control of neural plasticity. More broadly, the results should inform theories of visual recognition and perceptual learning, as well as clinical issues of rehabilitation and recovery from eye disease or injury.